Mojo function

batched_matmul_kernel_gpu

batched_matmul_kernel_gpu[c_type: DType, a_type: DType, b_type: DType, //, c_shape: DimList, a_shape: DimList, b_shape: DimList, transpose_b: Bool, config: MatmulConfig[a_type, b_type, c_type, transpose_b], elementwise_lambda_fn: OptionalReg[fn[DType, Int, Int, Int](IndexList[$2], SIMD[$0, $1]) capturing -> None] = OptionalReg[fn[DType, Int, Int, Int](IndexList[$2], SIMD[$0, $1]) capturing -> None]({:i1 0, 1})](c_buf: NDBuffer[c_type, 3, MutableAnyOrigin, c_shape], a_buf: NDBuffer[a_type, 3, MutableAnyOrigin, a_shape], b_buf: NDBuffer[b_type, 3, MutableAnyOrigin, b_shape], m: Int, n: Int, k: Int)